When searching for products or services online, a user is often presented with a large number of results. In an attempt to avoid overwhelming a user, data is generally organized in a manner that, hopefully, enhances the user experience. For example, the data is generally organized in manner that allows a user to quickly glean useful information from the presented search-query results.
Some example challenges that may exist with current techniques include the presentation of data in a meaningful manner and the expense of preparing data for the presentation. For example, data may be presented based on statistics that are user driven. However, presenting data solely based on user driven statistics may not be meaningful since user behavior is not constrained by a particular attribute of the product or service of interest. For instance, a user searching for a book about automotive repair may be presented with recommended books about cooking. Although user statistics may indicate that users often purchase these two types of books, there is no objective attribute-based reason for recommending a cooking book to a user looking at an automotive book. It may be that many user accounts are shared by married couples with diverse interests. However, such an underlying cause for data association is imperceptible to the user. To address this possible challenge, the book data may be tagged. However, determining tags and then tagging a vast database of data may require expenditure of significant resources.